/AyurGenomics-Viz-ML

In the Ayurveda system of medicine, individuals are classified into seven constitution types, “Prakriti”, for assessing disease susceptibility and drug responsiveness, which are assigned to an individual based on these phenotypes. We aim to develop machine learning algorithms for visualizing this heterogenous multi-dimensional phenomics and genomics data.

Primary LanguagePythonMIT LicenseMIT

AyurGenomics-Viz-ML

Ayurgenomics Visualisation and Machine Learning

Proposed title

'To develop machine learning algorithm for visualizing heterogenous multidimentional phenomics and genomics data'

Objectives

  • To develop ML algorithm to visualize individual-level signature based on multiple phenotypes

    • Develop algorithm to capture phenotype to phenotype relationship
    • Develop algorithm to capture within Prakriti signatures
    • Develop algorithm to capture between Prakriti signatures
    • Design visual representation of individual-level signature based on above algorithms

  • To develop ML algorithm to associate phenotype with gene expression

    • Use existing methodologies to find above associations
    • Develop novel ML algorithms to find multi-phenotype association with molecular cues.

BITS scholars invovled in the project

Mentors

TA

Background

Technological advancement in high-throughput experiments (HTE) allow us to decipher many biological insights such as, how transcription factor interact with downstream genes, with the aid of machine learning algorithms. Machine learning algorithm play a very vital and integral part of understand complex biological event where we profile multitude of genes and uncover patterns from it. Most HTE involve experients, where the phenotype of interest (Xpheno) is simple such as (case/control, normal/disease conditions) and accordingly we developed algorithms to infer genes(Yg) as predictors of the phenotypes (eg., cancer). In recent years, we started appreciating the fact that other covariate such as age, sex, environmental conditions along with our phenotype of interest could play a vital role in regulation within cellular. Overview of phenomics and genomics data is illustrated below in Figure-01.

What we did?

You may follow the link below for a detaiiled overview of the project.

Report

References